In the knowledge management field, knowledge map created under the client-server architecture has been widely used to direct the knowledge sharing process. Peer-to-peer (P2P) architecture has been practicable for file sharing, distributed computing, instant messaging, etc. by virtue of the increases of Internet bandwidth and personal computer capability. P2P architecture attracts researchers and practitioners to study knowledge sharing issues in its autonomy and self-organization characters. It calls for academic efforts to design knowledge management supporting system as that for client-server architecture may not be applicable. This study proposes a visualized cognitive knowledge map integration system to facilitate knowledge management on P2P networks. By using the SOM (Self-Organized Map)-like model, called Egocentric SOM (ESOM), the prototyping system can merge the external knowledge under a focal peer’s knowledge structure and present the cognitive knowledge map visually. In evaluating the proposed integration method, this study allocates abstracts of industrial research reports from an industrial technology research institute according to their corresponding author peers and generates individual cognitive knowledge maps. The results from the evaluation experiments reveal that ESOM is capable to retain individual peer’s knowledge structure while articulating with that of other peers in the cognitive knowledge network.